INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Slit sensilla are mechanoreceptors characteristic to spiders. They respond to strains that compress slits on the spider exoskeleton (Barth 1985). The lyriform VS-3 organ (nomenclature of Barth and Libera 1970) in the patella of the spider Cupiennius salei consists of seven to eight sensilla, each of which is innervated by a pair of mechanosensory neurons. Both neurons of each pair respond to a sustained mechanical or electrical stimulus with a rapidly adapting burst of action potentials. In type A neurons this adaptation takes place in <20 ms, often producing only one or two action potentials, but in type B neurons the response is a burst lasting several hundred milliseconds (Seyfarth and French 1994). Although this difference in adaptation properties can be partly explained by differences in the dynamic properties in the transduction step from the mechanical stimulus to the receptor potential (Juusola and French 1998), active membrane conductances seem to dominate the overall dynamic behavior.

Voltage-activated outward K⁺ currents have been shown to have an important function in regulating action potential firing frequency (Connor and Stevens 1971) as well as controlling the duration of individual action potentials (Hodgkin and Huxley 1952). The contribution of K⁺ currents to adaptation has been described in two different types of invertebrate mechanoreceptors: 1) the rapidly adapting tactile spine neuron of the cockroach (Torkkeli and French 1994, 1995), where an A current and a Ca²⁺-activated K⁺ current were shown to have major roles in the adaptation and the delayed rectifier current was the major action potential repolarizing current, and 2) the rapidly and slowly adapting stretch receptor neurons of the crayfish (Purali and Rydqvist 1992), where the voltage-activated outward currents were not shown to play major roles in adaptation. We examined the kinetics and pharmacological sensitivity of voltage-activated K⁺ currents in type A and type B neurons of the spider VS-3 organ to learn if these currents are responsible for the differences in their adaptation properties.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Preparation

A laboratory colony of Central American wandering spiders, Cupiennius salei, Keys., was kept at room temperature (22 ± 2°C). Legs from adult spiders of either sex were autotomized, and a concave piece of patellar cuticle containing slit sense organ VS-3 was dissected free in spider saline. The neurons are directly beneath the cuticle embedded in an internal membrane (hypodermis). We detached the hypodermis with the neurons attached to it from the cuticle as described by Höger and Seyfarth (1995) and placed them on a small round glass coverslip coated with 5 µg/ml Collagen IV (Sigma; Oakville, ON). The coverslip was fixed to the bottom of a 35-mm culture dish (Falcon 3001, Becton Dickinson; Franklin Lakes, NJ) (Fig. 1.). This "hypodermis preparation" differs from the previously described "cuticular preparation" of the VS-3 organ (e.g., Juusola and French 1995; Seyfarth and French 1994) in two important ways. First, the neurons are facing up, with the hypodermis below them, which makes it easier to impale the neurons with intracellular electrodes and allows faster exchange of solutions. Second, the axons and dendrites can be crushed to prevent current propagation and consequently improve voltage-clamp conditions. In these experiments the axons and dendrites were crushed at 100-200 µm from the somata. This preparation does not allow research into the mechanical properties of the VS-3 neurons because the mechanically activated ion channels are located on the tips of the dendrites (Höger et al. 1997). The action potential initiation zone obviously remains intact in this preparation because the neurons retain their typical adaptation patterns in response to depolarization.

View larger version (119K): [in this window] [in a new window]   Fig. 1. "Hypodermis preparation." Shaded area: hypodermis that covers the neurons in an intact preparation. Neurons appear in pairs, which are numbered from 1 to 7 from right to left. Type A and type B neurons cannot be identified visually because their locations with respect to each other vary in different preparations. Pair 2 is an exception because the type A neuron is always smaller than the type B neuron. The dendritic portion of the neurons is broken off from the cuticle, and the axons join the main leg nerves, which run to the CNS (proximal). Before the experiments the axons and dendrites were crushed at 100-200 µm from the somata.

Recording and stimulation

Current- and voltage-clamp recordings were performed with the discontinuous single-electrode method (Finkel and Redman 1984) with an SEC-10 l amplifier (NPI Electronic; Tamm, Germany). The conditions for successful single-electrode voltage and current clamp were described in detail by Torkkeli and French (1994), and the same methods were used previously in VS-3 neurons (Höger et al. 1997; Juusola et al. 1994). Borosilicate microelectrodes (1 mm OD and 0.5 mm ID) were pulled with a horizontal puller (P-2000, Sutter Instrument; Novato, CA). The electrodes were filled with 3 M KCl and coated with petroleum jelly to decrease stray capacitance (Juusola et al. 1997). Electrode resistances were 45-80 M with time constants of 1-3 µs in solution. Switching frequencies of 20-23 kHz and a duty cycle of 1/8 (current passing/voltage recording) were used in all experiments.

The neurons were observed with an inverted microscope with bright field optics (Olympus; Tokyo), identified, and then penetrated through a thin layer of spider saline. Penetration was achieved by high-frequency oscillation ("buzzing"). Impaled cells were allowed to stabilize for 15 min before the start of experiments. Criteria for reliable recordings were stable resting potentials of approximately 60 mV, action potential amplitudes of >40 mV, and thresholds for firing action potentials of 0.75 nA. With higher thresholds it was not possible to reliably classify type A and type B behavior because a small number of action potentials could indicate that the cell was damaged during penetration or that it normally fires only one spike.

All current- and voltage-clamp experiments were controlled by an IBM-compatible computer with custom written software (ASF Software; Halifax, NS). The computer provided voltage or current stimuli via a 12-bit D/A convertor. Membrane potential was low-pass filtered at 33.3 kHz and current at 3.3 kHz by the voltage-clamp amplifier. In current-clamp experiments the membrane resistance and time constant were estimated by measuring voltage responses to a hyperpolarizing current pulses of 0.25 nA. In voltage-clamp experiments the membrane resistance was calculated from the linear part of the current-voltage curve at hyperpolarizing potentials, and this value was used for leakage subtraction. Statistical analysis was made with the Student's two-tailed t-test for significantly different means, assuming different variances (Press et al. 1990).

Chemicals

The spider saline used in these experiments contained (in mM) 223 NaCl, 6.8 KCl, 8 CaCl₂, 5.1 MgCl₂, 5 mM sucrose, and 10 HEPES, pH 7.8 (Höger et al. 1997). Pharmacological agents used to block membrane currents were dissolved in spider saline and freshly prepared for each experiment or kept frozen at the same or higher concentrations. Voltage-activated Na⁺ currents were blocked with 1 µM TTX. In some experiments Ca²⁺ currents were blocked by replacing CaCl₂ with an equal concentration of CoCl₂, but the cells started to deteriorate rapidly in these conditions, and in most experiments the Ca²⁺ currents were blocked with 50 µM CdCl₂. Because Cd²⁺ is a more potent blocker of Ca²⁺ currents it can be used as an additive rather than replacement of all bath Ca²⁺ ions (Lancaster 1991). To keep the bath osmolarity unchanged when 5 mM tetraethylammonium chloride (TEA, Kodak) or 5 mM 4-aminopyridine (4-AP) were used, sucrose was removed from the saline. Ten to 100 nM charybdotoxin (CTX), 10-100 nM apamin, and 100 µM quinidine were added to the normal saline. The effects of blocking agents usually took place in 5-10 min. All chemicals were purchased from Sigma unless otherwise stated.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Passive membrane properties

Passive membrane properties of VS-3 neurons have been examined previously with the cuticular preparation (Juusola and French 1998). Here, we examined the passive membrane properties of neurons in the hypodermis preparation (Fig. 1) in which the axons and dendrites were crushed at 100- to 200-µm distance from the soma. The membrane resistance (R_in) and time constant () of both type A and type B neurons were significantly higher in the hypodermis preparation than reported for the cuticular preparation (Table 1). Mean values for type A neurons were R_in = 178.9 M and  = 16.8 ms (compared with 61.1 M and 6.3 ms, respectively, in Juusola and French 1998), and mean values for type B neurons were R_in = 114.1 M and  = 9.9 ms (70.8 M and 7.6 ms, respectively, in Juusola and French 1998). The differences between R_in and  values in type A and type B neurons were also statistically significant (Table 1). However, membrane capacitances were not significantly different between the neuron types (type A: 90.0 pF; type B: 81.9 pF) (Table 1). Because most of the neurons that we used for experiments had large somata (radius ~25 µm and length ~100 µm), the contributions of axons to membrane capacitance were quite small, and we found only a small difference in the membrane capacitance when compared with the cuticular preparation (110 pF) (Juusola and French 1998). Resting membrane potentials of type A and type B neurons were also close to each other (Table 1), suggesting similar resting conductances. We also compared the threshold depolarization, which was defined as the potential difference between the holding potential and the point where the action potential rose steeply (Fig. 2) (Nakajima and Onodera 1969). The mean threshold depolarization of type A neurons was 42.5 mV, and that of type B neurons was 36.5 mV. As shown in Table 1 this difference was statistically significant.

View larger version (15K): [in this window] [in a new window]   Fig. 2. Threshold depolarization was measured as the potential difference between the holding potential and the inflection point between the slowly rising potential and the steeply rising action potential. Stimulus amplitude in this recording was 500 pA. : mean threshold depolarization of type A neurons, 42.5 ± 8.9 mV (n = 33); : mean threshold depolarization of type B neurons, 36.5 ± 8.2 mV (n = 34). Difference was statistically significant (Table 1).

Voltage-activated outward currents

We recorded voltage-activated outward currents after inward Na⁺ currents were blocked with 1 µM TTX and Ca²⁺ currents were blocked with either Co²⁺ or Cd²⁺. Without Co²⁺ or Cd²⁺ in the bath, addition of K⁺-channel blockers (TEA, 4-AP, or quinidine) revealed a rapidly activating and inactivating inward current at high voltages. Blockade of Ca²⁺ currents would also inhibit Ca²⁺-activated K⁺ currents, but because the amplitude of the outward currents did not change significantly and in a few experiments it actually increased after administration of Ca²⁺-channel blockers rather than decreased we do not believe that a large fraction of outward current was dependent on Ca²⁺.

Blockade of inward currents increased the membrane resistance approximately twofold; type A neurons had a mean membrane resistance of 395.6 M, and type B neurons had a resistance of 240.3 M. This difference in the membrane resistances was statistically significant (Table 1). Both types of VS-3 neurons had a similar profile of outward currents (e.g., Fig. 7A, type A, and Fig. 8A, type B) with a rapid activation and slower inactivation at high recording voltages. These currents could be separated into two parts with the pulse protocols shown in Fig. 3.

View larger version (22K): [in this window] [in a new window]   Fig. 3. Voltage-clamp recordings after inward currents were blocked with TTX and Cd2+. Actual voltage records are shown below each set of current recordings (A-C). A: slightly hyperpolarizing prepulses were used to elicit transient outward current. B: noninactivating currents produced with a depolarizing prepulse were subtracted from the currents in A. C: inactivation kinetics of the transient outward current were studied with prepulses to different voltages and by recording the transient at a constant level. D: transient outward current after subtraction of the trace at 30 mV in B from the corresponding trace in A. Current trace was fitted by 2 exponential decay time constants (Eq. 5).

Transient outward current

Rapidly activating and inactivating outward current could be elicited with a stimulus from resting or hyperpolarized potential (Fig. 3A) and the transient component revealed by subtracting a current that was produced after a depolarizing prepulse (Fig. 3B). The maximum amplitude of the transient outward current was 0.24 nA in type A neurons and 0.45 nA in type B neurons. Voltage dependencies of activation and inactivation of transient outward current were studied by measuring the peak current during double-voltage pulse protocols (Fig. 3, A and C). The experimental activation data were fitted by a Boltzmann distribution of the form

(1)

where g is conductance, V is membrane potential, V₅₀ is the membrane potential at half-maximal activation, and s is the slope factor. Conductance was calculated from

(2)

where I is the current and E_rev is its reversal potential. Fitting was performed by the Levenberg-Marquardt general nonlinear fitting algorithm (Press et al. 1990). The reversal potentials were defined from the tail currents of the peak transient outward currents (Ritchie 1987; Torkkeli and French 1995). The results obtained from fits of 11 type A neurons and 10 type B neurons are shown in Table 2, and Fig. 4 shows examples of Boltzmann fits in both types of neurons and bar graphs of V₅₀ values from all recordings. V₅₀ values of type A neurons were significantly more positive than type B neurons, but all other parameters were similar in both neuron types.

View larger version (14K): [in this window] [in a new window]   Fig. 4. Voltage dependence of activation of the peak transient outward current. : fitted Boltzmann distributions to normalized conductances. : data from type A neuron; : data from type B neuron. Bar graphs: V50 values for activation data from 11 type A and 10 type B neurons. Mean V50 value of type A neurons was 9.6 ± 16.1 mV, and of type B neurons it was 13.1 ± 15.1 mV. Difference was statistically significant (Table 2).

The equivalent gating charges (e) for activation was calculated from

(3)

where k is the Boltzmann constant, T is absolute temperature, e is the elementary charge, and z is the valency of equivalent gating charge. For type A neurons with s = 12.6 mV (Table 2) and kT/e = 25.26 mV (Hille 1992) the gating charge for activation was 2.0, and for type B neurons with s value of 14.1 mV (Table 2) it was 1.8. The activation in both cases was clearly voltage dependent and followed a sigmoidal time course that indicates multiple gating particles (Hodgkin and Huxley 1952).

Equations 1 and 2 were also used to fit the inactivation data with appropriate change in exponential sign. The fitted parameters from 8 experiments with type A neurons and from 10 experiments with type B neurons are given in Table 2. One example of a Boltzmann fit of each neuron type is shown in Fig. 5. A similar shift in the V₅₀ values was visible in the inactivation as in the activation, but this difference was not statistically significant. The amplitude of transient outward current did not increase when prepulse potentials of less negative than 70 mV were used, indicating that this current was not the typical A current found in many other neurons.

View larger version (16K): [in this window] [in a new window]   Fig. 5. Voltage dependence of inactivation of the peak transient outward current. : fitted Boltzmann distributions to normalized conductance. : data points from 1 type A neuron; : data points from 1 type B neuron. Bar graphs: V50 values for inactivation data from 8 type A neurons and 10 type B neurons. Mean V50 values were 38.4 ± 12.0 mV for type A neurons and 31.1 ± 9.2 mV for type B neurons. Difference was not statistically significant (Table 2).

The time course of the transient outward current was fitted with exponential power equations with either one or two inactivation time constants (Eqs. 4 and 5). Two time constants of inactivation were needed when the current amplitude was larger.

(4)

(5)

where I is the current level expected in the absence of inactivation, ₁ is the time constant of activation, ₂ and ₃ are time constants of inactivation, m is the integer exponent, and  is the relative amplitude of the two inactivation components (Hodgkin and Huxley 1952). An example of a fit with Eq. 5 with an m value of 4 is shown in Fig. 3D. The activation and inactivation time constants of transient outward current are given in Table 2, and there were no statistically significant differences between type A and type B neurons.

Noninactivating outward current

Voltage sensitivity of activation of the noninactivating outward current was studied with the steady-state currents from recordings such as Fig. 3B, where a depolarizing prepulse was used to eliminate the transient outward current. Noninactivating current made up a major part of the outward current. Its mean maximum amplitude was 1.2 nA in type A neurons and 1.7 nA in type B neurons (Table 3). The steady-state activation was well fitted by the Boltzmann distribution (Eq. 1; Fig. 6). The reversal potential was defined from the tail currents of the steady-state currents after 100-ms pulses to several depolarizing voltages (Ritchie 1987; Torkkeli and French 1995). The fitted parameters from 17 sets of data from type A neurons and 18 recordings of type B neurons are shown in Table 3. The only differences that were statistically significant between the two neuron types were positive shifts in the V₅₀ values of the type A neurons when compared with type B neurons that was also found in the transient outward current, and also the s value was significantly more positive in type A neurons than in type B neurons (Table 3). It was not possible to study the time course of the noninactivating outward current because the current was already activated at the start of the recording. The equivalent gating charge (e) for the activation of noninactivating outward current was calculated with Eq. 3. For type A neurons it was 2.3, and for type B neurons it was 3.2. 

View larger version (16K): [in this window] [in a new window]   Fig. 6. Voltage dependence of activation of noninactivating outward current. : fitted Boltzmann distributions to normalized conductances. : data points from a type A neuron; : data points from a type B neuron. Bar graphs: mean values of V50 from 17 type A neurons (48.9 ± 4.9 mV) and 18 type B neurons (56 ± 4.2). Difference was statistically significant (Table 3).

From mean reversal potentials and the value of 6.8 mM for extracellular K⁺ concentration we calculated with the Nernst equation that the intracellular K⁺ concentration was 199 mM in the type A neurons and 190 mM in the type B neurons.

Pharmacology

Experiments with the potassium channel blockers TEA and 4-AP are shown in Figs. 7 and 8. When 5 mM TEA was added to the bath solution it blocked the transient outward current and a small part of the noninactivating outward current. The experiment shown in Fig. 7 was done in a type A neuron, and results were very similar with type B neurons. 4-AP (5 mM) also blocked the transient current and almost one-half of the noninactivating current (Fig. 8). The blocking effect of 4-AP was stronger than that of TEA in both neuron types. Larger concentrations of either blocker (25 and 50 mM) did not produce stronger block. Similar effects were produced when 100 µM quinidine was used as a blocker. We also tried two toxins that block Ca²⁺-sensitive K⁺ currents in many neuron types, apamin and CTX (Hille 1992), both at 10- to 100-nM concentrations. These experiments were done in normal bath Ca²⁺ concentrations and without Co²⁺ or Cd²⁺ in the saline. There were no differences in the outward current amplitudes before or after applying the blockers.

View larger version (33K): [in this window] [in a new window]   Fig. 7. Tetraethylammonium chloride (TEA) blocked the transient outward current and a small portion of the steady outward current in this type A neuron. The pulse protocol below in A is the actual voltage recording from the amplifier. Leakage currents were subtracted from both current recordings. B: current recordings from a 30-mV voltage pulse before (Control) and after TEA in the bath together with the current that was blocked by TEA (Subtracted).

View larger version (32K): [in this window] [in a new window]   Fig. 8. Effect of 4-aminopyridine (4-AP) on the outward currents in a type B neuron. Actual voltage from the amplifier is shown below in A and the current recordings with (4-AP) and without 4-AP (Control) above. B: when current after 4-AP application was subtracted from the control current it revealed a block of the transient outward current and ~1/2 the steady current (Subtracted).

The effects of K⁺-channel blockers on the action potentials were studied in the current-clamp configuration (Fig. 9). The effects of 4-AP and TEA were very similar; the amplitude and duration of action potentials increased, but the firing frequency was not affected in either neuron type. Type A neurons fired usually one large action potential after 4-AP or TEA was added to the bath solution (Fig. 9A), and type B neurons produced smaller numbers of action potentials after the blockers (Fig. 9B), but this occurred because of longer-duration spikes rather than a direct effect on firing frequency. The current that was needed to induce action potentials in both neuron types was reduced by both blockers, and in some experiments they produced small spontaneous depolarizations. For example, in the recording shown in Fig. 9A, the minimum stimulus current amplitude to produce an action potential fell from 500 to 250 pA after 4-AP application. In the recording shown in Fig. 9B the current amplitude needed to cross the threshold decreased from 750 to 250 pA. The effect of quinidine was stronger than that of 4-AP or TEA. It actually prevented the neuron from repolarizing after the first action potential (Fig. 9C). Ten to 100 nM apamin or CTX did not affect individual action potentials or firing patterns.

View larger version (13K): [in this window] [in a new window]   Fig. 9. Effects of K+-channel blockers on action potentials. Amplitude and duration of action potentials increased when 4-AP was added to the bath of a type A neuron (A). Stimulus amplitude in control and 4-AP recordings was 500 pA. When TEA was added to the bath of a type B neuron (B) the amplitude and duration of action potentials increased as well as the baseline depolarization, indicating that the repolarization was partially inhibited. Stimulus amplitude in control recording was 1.25 nA, and in TEA recording it was 1 nA. C: quinidine completely prevented cell repolarization after the first action potential in a type B neuron. Stimulus amplitude in control and quinidine recordings was 750 pA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Since the cuticular preparation of the spider VS-3 organ was first developed by Seyfarth and French (1994), it has been successfully used to study responses of the sensory neurons to mechanical stimuli (e.g., Höger et al. 1997; Juusola et al. 1994). This preparation provides a useful tool for such studies even in the voltage-clamp configuration because the voltage control is good enough to examine small amplitude current changes produced by mechanical stimuli when rapidly changing voltage-activated currents are blocked (Juusola and French 1998). However, the control of voltage-activated currents can be greatly improved when their propagation into the neural processes is inhibited. Therefore we utilized here the hypodermis preparation developed by Höger and Seyfarth (1995), which allowed crushing the neurites at a short distance from the soma. Another benefit of this preparation was improved access to the neurons with pharmacological agents. The passive membrane properties of the neurons were significantly different when the results from the hypodermis preparation were compared with the cuticular preparation (Juusola and French 1998), with higher membrane resistances and longer time constants indicating that the numbers of open ion channels were smaller in the hypodermis preparation. This result is expected when the propagation of currents into the processes is prevented. The differences in membrane resistances and time constants between type A and type B neurons were new findings (Table 1). The membrane capacitances in the type A and type B neurons were not significantly different, indicating that the average sizes of these neurons are not different, although a clear size difference has been found in one pair (pair 2, Fig. 1), where the type A neuron is always smaller than the type B neuron (Fabian-Fine et al. 1999). These results indicate that the resting conductance of the type B neuron is larger than the type A neuron, at least on the soma and the parts of neurites close to the soma. The threshold depolarization was higher and voltage-activated outward currents activated at more depolarizing voltages in type A neurons than in type B neurons, indicating differences in the relative distributions and densities of both the Na⁺ and K⁺ channels in these neurons.

Although the voltage-activated outward currents in VS-3 neurons could be separated into two components, these components did not clearly fall into the two major categories of voltage-activated outward currents that have been found in many other types of neurons in invertebrates and vertebrates, namely the A current and delayed rectifier (Hille 1992). However, there were similarities. The noninactivating outward current probably has the same function in repolarizing the action potentials as the delayed rectifier, although it activated without delay and did not have strong sensitivity to TEA. The transient outward current in the VS-3 neurons activated maximally from rest and did not need a hyperpolarizing prepulse to be fully activated, which is different to the classical A current (Connor and Stevens 1971; Rogawski 1985; Torkkeli and French 1994). However, it had a very fast time course, and it may have functions similar to the A current. From the firing patterns of VS-3 neurons (Fig. 9) it can be seen that a hyperpolarization-induced A current cannot be involved because the action potentials return to a depolarized level as long as the stimulus is applied and do not hyperpolarize below the resting potential. However, they return to a level where the transient outward current in these neurons starts to activate. In both neuron types the number of action potentials increased with increasing stimulus amplitude, although in type A neurons even the largest depolarizations did not produce more than two or three action potentials. The function of encoding graded depolarization into firing frequency is believed to depend strongly on A current because some neurons and the axons of many neurons that do not have an A current fire similar bursts of action potentials independent of stimulus intensity (Rogawski 1985). The transient outward current in VS-3 neurons could be activated at the range of voltages where it can affect frequency encoding, and, although the K⁺ channel blockers that we used were not specific for the transient outward current, they usually decreased the threshold for firing action potentials as well as increasing the duration and amplitude of the spikes. None of the blockers produced an increase in the firing frequency, but if a type A neuron originally had more than one action potential the blockers usually reduced the number to one, and they decreased the rate of repetitive firing in type B neurons. These effects indicate that neither of these currents could be responsible for action potential adaptation because removal by blockers would then induce more vigorous firing. The action potential repolarizing function of the noninactivating outward current in VS-3 neurons is similar to that of the delayed rectifier K⁺ current in many neurons of other invertebrate species such as molluscan soma (Adams et al. 1980; Aldrich et al. 1979), Aplysia bursting pacemaker neurons (Adams and Benson 1985), and cockroach tactile spine neurons (Torkkeli and French 1995). In vertebrates this current is less important in spike repolarization (Storm 1987). For example, in frog sympathetic ganglia the Ca²⁺-activated K⁺ current is the main repolarizing current (Adams et al. 1980), and in rat ganglia the A current has an important function in action potential repolarization (Beluzzi et al. 1985).

We found only minor differences between type A and type B neurons. Both had transient outward currents with small maximum conductances that activated and inactivated at approximately the same rate in both neurons (Table 2). Also the membrane conductance responsible for the noninactivating current was similar in type A and type B neurons (Table 3). The only significant differences were in the activation kinetics of outward currents because both of them activated at more positive potentials in type A neurons than in type B neurons (Tables 2 and 3).

We did not find any evidence for the existence of Ca²⁺-activated K⁺ current in the VS-3 neurons. Experiments with CTX or apamin did not reveal sensitivity to these blockers in current- or voltage-clamp recordings. However, it is possible that these neurons have K⁺ current that is activated by Ca²⁺ but not sensitive to these toxins. Purali and Rydqvist (1992) did not find a Ca²⁺-activated K⁺ current in the slowly or rapidly adapting stretch receptor neurons of crayfish using the same blockers, but single-channel recordings revealed that Ca²⁺-activated K⁺ channels are present in these neurons (Erxleben et al. 1991). The density of these channels in some neurons could be so small that the whole cell currents cannot be easily detected. Ca²⁺-activated K⁺- current was found to be one of the main components of adaptation in one type of insect cuticular mechanoreceptor neuron, the cockroach tactile spine (Torkkeli and French 1995), where approximately one-half of the total outward current was activated by Ca²⁺. The mechanisms of adaptation of VS-3 neurons are probably very different from those of the tactile spine because their adaptation behavior is different. The tactile spine is a phasic receptor, but it can fire bursts that last several seconds (Torkkeli and French 1995), the action potentials always return to hyperpolarizing potentials, and their threshold depolarization is substantially lower than VS-3 neurons (~5 mV, unpublished observation). Consequently, the tactile spine neuron has a large A current that activates when the membrane is hyperpolarized below the resting potential (Torkkeli and French 1994). The VS-3 neurons do not have the slow afterhyperpolarization that has been shown to be produced by Ca²⁺-activated K⁺ currents in many neurons and is assumed to slow the rate of firing (Hille 1992).

In investigations of voltage-activated outward currents of slowly and rapidly adapting crayfish stretch receptor neurons (Rydqvist and Purali 1991; Rydqvist and Zhou 1989) the activation of outward currents was faster and occurred at more negative potentials in the rapidly than slowly adapting neurons. These findings are opposite to our results from VS-3 neurons, but the adaptation behavior of stretch receptor neurons is also very different from VS-3 neurons because both the rapidly and slowly adapting neurons can fire bursts of action potentials several seconds long, and VS-3 neurons only fire for a maximum of a few hundred milliseconds. However, Purali and Rydqvist (1992) did not regard the differences in the outward current kinetics as a major contributor to the variations in the adaptation behavior of stretch receptor neurons. Likewise, we do not believe that the small differences we found here between the activation kinetics of type A and type B neurons can have a major effect on the adaptational differences of VS-3 neurons. It seems more likely that the firing pattern is defined by other factors, possibly differences in the inactivation kinetics of the inward currents. Recently, sodium current inactivation has been shown to begin at more negative potentials, and it is also faster in the rapidly than in the slowly adapting stretch receptor neuron of the crayfish (Purali and Rydqvist 1998), suggesting a major regulatory function. Neither sodium nor calcium currents of VS-3 neurons have yet been investigated, and these currents may dominate their dynamic properties.